MediaCapture

MediaCapture is a simple, production-ready Python utility for screen and camera capture, screen/camera streaming, multi-camera compositing, and media normalization—using only a few lines of code.

Features

• Capture screenshots to file
• Record screen to video
• Capture images from one or more cameras (with optional normalization)
• Record video from multiple cameras to a single file (side-by-side)
• Real-time screen streaming (with auto-resize)
• Real-time multi-camera streaming (with auto-resize, optional normalization)
• Simple directory management
• No external dependencies beyond common imaging libraries

Installation

pip install MediaCapture

Usage

from MediaCapture import MediaCapture
import cv2
import numpy as np

mediaCapture = MediaCapture()

# Capture a screenshot
mediaCapture.captureScreen('screenshots/screen.jpg')

# Record the screen for 10 seconds
mediaCapture.recordScreen('videos/screen.mp4', duration=10)

# Capture from two cameras and save side-by-side
mediaCapture.captureMedia('images/cameras.jpg', 0, 1, normalize=True)

# Record from two cameras for 15 seconds
mediaCapture.recordMedia('videos/cameras.mp4', 0, 1, duration=15, normalize=True)

# Stream live screen preview (press 'q' to exit)
for frame in mediaCapture.streamScreen(fps=15):
    cv2.imshow("Screen", cv2.cvtColor(frame, cv2.COLOR_RGB2BGR))
    if cv2.waitKey(1) & 0xFF == ord('q'):
        break
cv2.destroyAllWindows()

# Stream from cameras (press 'q' to exit)
for frames, frameCount in mediaCapture.streamCameras(0, 1, normalize=True):
    if any(f is None for f in frames):
        break
    combined = np.hstack(frames)
    cv2.imshow("Combined Cameras", combined)
    if cv2.waitKey(1) & 0xFF == ord('q'):
        break
cv2.destroyAllWindows()

Methods

Notes

• ImageGrab requires Windows/macOS or X11-based Linux.
• For multi-camera capture, you must have multiple camera devices attached.
• Streaming windows can be exited by pressing the q key.
• Output directories are created automatically if they don’t exist.

License

This project is licensed under the Apache License, Version 2.0. Copyright 2025 Tristan McBride Sr.

Acknowledgements

Project by:

• Tristan McBride Sr.
• Sybil